In 2012, we implemented the Curriculum and Assessment Policy Statements (Caps) for Grades 1-3 and Grade 10. Those Grade 1 students are 19-20 now, finishing university or entering the workforce. Those Grade 10 students are 26-27, starting their careers.
And everywhere I turn, I hear the same complaints about this generation: “They can’t think critically.” “They just want to be told the answer.” “They lack problem-solving skills.”
We’re blaming smartphones. We’re blaming ChatGPT. We’re blaming TikTok and “kids these days”.
But we’re not blaming the curriculum that taught them compliance over inquiry. We’re not blaming the assessment system that rewarded memorisation over understanding. We’re not blaming the pedagogical approach that punished questioning.
This is the 15-20 year harvest. The seeds planted in 2012 are being reaped in 2025.
Strategising while schools collapse
The Department of Basic Education spent seven years developing the Funda Uphumelele National Survey – sophisticated assessment tools that can measure with precision exactly how our children are failing to read across all 11 official languages. Seven years of linguistic analysis, benchmark creation, tool validation and pilot testing.
You know what we didn’t do in those seven years? Ensure that teachers can actually teach reading in those 11 languages.
Most South African teachers are English or Afrikaans speakers. Only about 15% can assess – let alone teach – in African languages. We created thermometers calibrated to measure temperature in 11 different rooms. But we never built the heaters.
Call this the instructional void – the gap between our capacity to measure failure and our capacity to prevent it.
The design reveals the priorities
I pulled up the Foundation Phase Caps documents recently – Grades R-3, where literacy foundations get built or don’t.
The English Home Language document: 142 pages.
The Mathematics document: Triple that size. Much more detail. Many more exemplars. Significantly more teacher guidance.
Mathematics – which depends entirely on literacy as its foundation – gets triple the instructional support compared with the language document itself.
This isn’t accidental. This is design revealing priorities.
The cliff drop
Then there’s the language transition. Caps mandates mother-tongue instruction for Grades 1-3, then switches to English or Afrikaans in Grade 4. Not gradually. Not with scaffolding. Just off the cliff.
Compare this to Sweden, where immigrant children learn Swedish while simultaneously learning content in their home language. There’s no separation of skill and language acquisition – they happen in parallel. By the time Swedish becomes the primary language of instruction, students are already competent in it.
South Africa’s approach looks right on paper – mother-tongue education! – but functions in reverse of its stated goals. It creates what materials scientists call chirality – a mirror image that can’t be superimposed on to what actually works.
What the Unesco report won’t tell you
The recently released Unesco Global Education Monitoring Report identifies critical structural problems driving South Africa’s literacy crisis. The data is damning. The recommendations are comprehensive: teacher training, resource allocation, multilingual instruction support and administrative capacity building.
All necessary. All correct. All insufficient.
But, here is what the report can’t say: these solutions assume the system wants to fix the problem. They assume good-faith incompetence rather than structural design.
When a system spends seven years developing sophisticated measurement tools while teachers can’t teach in the languages being measured, that’s not incompetence. When curriculum documents give triple the support to subjects that depend on literacy compared with literacy itself, that’s not oversight. When language transitions are designed to fail students at the exact moment academic demands increase, that’s not an accident.
These are features, not bugs.
What actually works: Starting with dignity
John Dewey understood something most education reformers forget: you can’t teach democratic citizenship to children whose basic autonomy you don’t respect. You can’t build internal locus of control in children whose bodily needs you ignore. You can’t develop critical thinking in students whose questions you punish.
Here’s a practical example that sounds trivial, but isn’t: Let children go to the toilet when they need to.
Not “wait for break”. Not “you should have gone earlier”. Not “only two bathroom passes per day”. When a child’s body signals a need, that child should be trusted to respond.
This teaches children that:
- Their internal signals are valid and trustworthy;
- They have ownership over their own bodies;
- They possess an internal locus of control; and
- Other people have the same autonomy they do.
These aren’t soft skills. These are the cognitive foundations that make literacy instruction work. A child who has learnt to trust their physiological signals, who knows they have agency over their own body, who understands that autonomy is mutual, is a child who is cognitively ready to engage with complex text in ways that a compliance-trained child is not.
The reality that makes this recommendation sound naive to teachers on the ground: Bela’s class-sized regulations have stacked the system against these odds. When you have 40 children in a classroom, autonomous bathroom breaks aren’t a pedagogical choice – they become a safety liability. You can’t maintain the duty of care. You can’t track who’s where. One child leaving triggers five more asking. The classroom becomes ungovernable.
So teachers who philosophically believe in respecting student autonomy find themselves forced to implement authoritarian control policies because the structural conditions make autonomy dangerous. They’re reduced to crowd control and administrative compliance, not because they lack vision, but because the system has made vision impossible to implement.
This is how extraction works: Create conditions that require compliance-based control, then point to that control as evidence that teachers don’t value student autonomy. The system demands crowd management, then blames teachers for managing crowds.
Unesco’s recommendations focus on instructional technique, language support and resource allocation. These matter. But they sidestep the deeper question: What happens when we design systems that treat children as autonomous beings rather than compliance machines? And more critically: What happens when regulations like Bela make it structurally impossible to implement that autonomy, even when teachers desperately want to?
The harvest ahead
The 2012 Caps cohort is entering the workforce now. We’re seeing the results: young adults conditioned for format compliance rather than substance verification, trained to wait for explicit instruction rather than exercise judgement, rewarded for performing learning rather than actually learning.
Meanwhile, we’re planting seeds for 2040. Those seeds look remarkably similar to those we planted in 2012: more assessment, more measurement, more strategic planning. More thermometers for rooms without heat.
The harvest matches the seeds. Every single time.
If we want different outcomes in 2040, we need to plant different seeds in 2025. That means starting with a fundamental question the Unesco report can’t ask: Are we designing systems that tend human capacity, or systems that mine human compliance?
Because the literacy crisis won’t be solved by better measurement of failure. It’ll be solved by building systems that respect children’s dignity enough to let their capacity flourish.
And that starts with things as simple – and as profound – as trusting a child who needs to pee. DM